Ben Myers: Statement on Generative AI

https://benmyers.dev/ai-statement/

by jrepinc • 3 years ago

25 11 3 years ago

benmyers.dev

davedx• 3 years ago

Sure is a lot of negativity and cynicism around at the moment.

Tools and tech aren’t bad, the way people choose to use them might be.

SantalBlush• 3 years ago

Since it's impossible to prevent all misuse of a specific tool or piece of tech, the question is whether it brings a net benefit to society in spite of that misuse. If the answer is "no", then it's perfectly reasonable to call it bad.

narrator• 3 years ago

I don't like the thing because of small flaw x that negates overwhelming benefit to humanity y. No viable x will ever be acceptable to me because it will always have the capability to have flaw x. I will always focus on small negative x. The only solution is to elevate someone to power and let them ignore the constitution to get rid of the thing and prevents all small flaw x forever through continued increases in their power no matter what the voters say, and no matter how much they like y.

Meanwhile, other countries where they have the thing must also be stopped from doing so by international treaty or even war. A global government must be given unlimited power to stop the thing everywhere. Let's campaign for a global government with unlimited power to stop all small flaws y that we find in society, even though that's unpopular and the thing is democratically popular.

king_magic• 3 years ago

Ah, the holier than thou purists come out to play. The hypocrisy is pretty rich.

thiago_fm• 3 years ago

That article about OpenAI exploitation of kenyan workers hits me hard.

OpenAI and most AI companies ahead of their pack can only exist in the American wild west.

Also, current generative AI models is more akin to a compression algorithm than real intelligence (that can invent and discover stuff...). I don't believe so many people drank the kool-aid from Sam Altman. Much worse is seeing people that spent their whole life in the academia unable to discern this silliness and jumping on this bandwagon.

I guess everyone wants that $1 million salary from OpenAI, or that NVidia stocks go high.

Serious countries with real data protection laws would make those companies either cease to exist, or take many more years to get to where they are.

It won't take long(maybe a year or so) until people understand that they are building not an "AI" or intelligence, but an atomic bomb.

We need this though. This will enlighten ourselves more about technology and it's uses. It will also completely destroy most big techs, as they're all about collecting people's data.

I hope this is the catalyst that the congress needs to explode Alphabet, Meta and so on.

Lemmi2• 3 years ago

Apparently the workers were still well paid.

But this exploitation is everywhere.

The resources of my car have been extracted from another country and only a handful profit from that.

cloogshicer• 3 years ago

I like what he wrote at the very bottom:

... this is fallaciously anthropomorphizing machine learning, reasoning about the model as though it were sort of like a new consciousness that just blinked into existence, absent the context of its creation. Instead, I'd offer an alternative mental model: generative AIs are tools, devised and deployed by corporations to operate at scale, laundering content the corporations generally do not own with the active intention of reducing the need for, and the value of, human labor.

Lemmi2• 3 years ago

Which is still a good thing.

The transition to a Star Trek society is not doable otherwise.

chpatrick• 3 years ago

Can't it be both?

ss1996• 3 years ago

A big pet peeve of mine are articles that mention salary in US dollars without any context. It's common sense that $2 in Kenya goes much further than $2 in the US.

Doing some cursory googling - Kenya's median salary is around $900 monthly, while USA's is a little more than $4000. So the $2 figure in the Time article is actually $8+ for a Kenyan, which is higher than USA's federal minimum wage.

Alternatively, we could also normalise by Purchasing Power Parity between USA and Kenya, which is around 3.12, so $2 becomes $6+

Are these figures low? Sure. But exploitative? Don't think so.

myoffe• 3 years ago

Weak arguments. Labor exploitation is nothing special in regards to AI, neither is environmental impact. It's like everything else. You can say the same thing about Uber, Zara, Supermarkets, any tech company. This isn't really a valid argument against AI. And whether it's real intelligence or not is just ridiculous. Nobody cares. It's useful, it's all that matters.

Ben Myers: Statement on Generative AI

Related Stories